Edit model card

NER model based on allenai/scibert_scivocab_cased Fine-tuned using the SciERC Dataset to identify scientific terms:

  • Task: Applications, problems to solve, systems to construct. E.g. information extraction, machine reading system, image segmentation, etc.
  • Method: Methods , models, systems to use, or tools, components of a system, frameworks. E.g. language model, CORENLP, POS parser, kernel method, etc. • Evaluation Metric: Metrics, measures, or entities that can express the quality of a system/method. E.g. F1, BLEU, Precision, Recall, ROC curve, mean reciprocal rank, mean-squared error, robustness, time complexity, etc.
  • Material: Data, datasets, resources, Corpus, Knowledge base. E.g. image data, speech data, stereo images, bilingual dictionary, paraphrased questions, CoNLL, Panntreebank, WordNet, Wikipedia, etc.
  • Other Scientific Terms: Phrases that are scientific terms but do not fall into any of the above classes E.g. physical or geometric constraints, qualitative prior knowledge, discourse structure, syntactic rule, discourse structure, tree, node, tree kernel, features, noise, criteria
  • Generic: General terms or pronouns that may refer to an entity but are not themselves informative, often used as connection words. E.g model, approach, prior knowledge, them, it...

Training

  • Learning Rate: 1e-05
  • Epochs: 10,

Performance

  • Eval Loss: 0.401
  • Precision 0.577
  • Recall: 0.632
  • F1: 0.603

Colab

Check out how this model is used for NER-enhanced topic modelling, inspired by BERTopic.

Use

from transformers import AutoTokenizer, AutoModelForTokenClassification
from transformers import pipeline

tokenizer = AutoTokenizer.from_pretrained("RJuro/SciNERTopic")
model_trf = AutoModelForTokenClassification.from_pretrained("RJuro/SciNERTopic")

nlp = pipeline("ner", model=model_trf, tokenizer=tokenizer, aggregation_strategy='average')

Cite this model

@misc {roman_jurowetzki_2022,
    author       = { {Roman Jurowetzki, Hamid Bekamiri} },
    title        = { SciNERTopic - NER enhanced transformer-based topic modelling for scientific text },
    year         = 2022,
    url          = { https://huggingface.co/RJuro/SciNERTopic },
    doi          = { 10.57967/hf/0095 },
    publisher    = { Hugging Face }
}
Downloads last month
150
Safetensors
Model size
109M params
Tensor type
I64
·
F32
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.